Traffic monitoring device, traffic monitoring system and method

Abstract

The invention relates to a traffic monitoring device (10) with at least one detection unit (12) for detecting the speed of at least one vehicle (14, 16) on a roadway (18) based on relative changes in light intensity generated by the vehicle (14, 16). To provide a generic device with improved efficiency, it is proposed that the detection unit (12) comprise at least one event-based image sensor (20) comprising a pixel matrix (22) with a plurality of pixels (24, 26), wherein the pixels (24, 26) are each configured to detect relative changes in light intensity independently and asynchronously as events (28, 30).

Description

Technical field

[0001] The invention relates to a traffic monitoring device according to the preamble of claim 1, a traffic monitoring system according to claim 10, and a method for determining the speed of at least one vehicle according to the preamble of claim 11. State of the art

[0002] Driving at excessive speed is one of the most frequent causes of accidents. Traffic monitoring devices can therefore contribute to improving road safety. A variety of such devices are already known from the prior art, based on different technologies, such as radar, lidar, or induction loops. Traffic monitoring devices based on various optical sensors are also known. For example, EP 0 042 546 A1 discloses a speed measuring system with three reflective barriers arranged at a specific distance from each other in the direction of movement of a measured object, which are generated by active light sources.Furthermore, so-called single-sided sensors are known from the prior art. These sensors detect the speed of vehicles based on relative changes in light intensity as the vehicle passes by, using passive contrast diodes. Each diode, equipped with its own lens, is spaced apart in a row and oriented perpendicular to the direction of travel. This eliminates the need for an active light source. Disadvantages of these previously known traffic monitoring devices based on optical sensors include the large space requirement for the spaced-apart active reflection barriers or passive contrast diodes, as well as the associated high costs.Another disadvantage of these known traffic monitoring devices is that the active reflective barriers must be aligned very precisely, at the same distance and angle to the roadway. This regularly leads to operator errors and thus to the rejection of measurements or an increase in the tolerance of measured speeds. With known traffic monitoring devices using passive contrast diodes, this disadvantage is less pronounced; however, the measurement performance suffers from insufficient alignment accuracy, as speeds are underestimated. Furthermore, the accuracy is limited by the spacing between the individual contrast diodes, which is restricted by the housing.

[0003] The object of the invention is, in particular, to provide a generic device and a generic method with improved efficiency. This object is achieved according to the invention by the features of claims 1, 10, and 11, while advantageous embodiments and further developments of the invention can be found in the dependent claims. Description of the invention

[0004] The problem is solved by the characteristics of independent claims.

[0005] The invention relates to a traffic monitoring device with at least one detection unit for measuring the speed of at least one vehicle on a roadway based on relative changes in light intensity generated by the vehicle.

[0006] It is proposed that the detection unit includes at least one event-based image sensor comprising a pixel matrix with a plurality of pixels, each pixel being configured to detect relative changes in light intensity independently and asynchronously as events.

[0007] This design allows for the advantageous provision of a particularly efficient traffic monitoring device. Material and installation space efficiency can be improved by significantly reducing the number of components, such as lenses, compared to previously known solutions. This results in a particularly compact traffic monitoring device. Furthermore, cost efficiency can be improved, leading to a particularly inexpensive traffic monitoring device. Since the pixels of the event-based image sensor that are not triggered by changes in light intensity are not activated, no events are recorded in static scenes. This enables particularly energy-efficient operation of the traffic monitoring device with comparatively low data volumes.Furthermore, adjustment can be significantly simplified if the detection unit has at least one event-based image sensor, since, compared to the previously required alignment of multiple sensors, now only the alignment of a single event-based image sensor to the road is sufficient to enable precise and reliable vehicle speed measurement. Horizontal deviations can even be digitally adjusted afterward. Due to the high pixel density, the horizontal alignment relative to the road is practically irrelevant. Even if, in extreme cases, a vehicle moves diagonally across the event-based image sensor, its speed can be determined by appropriately calibrating the actual horizontal plane of the image area. The actual horizontal plane can be determined, for example, from the direction of movement of a temporal sequence of detected events.This is possible with prior calibration, but can also be done individually for each vehicle detected. This results in significantly improved ease of use and a considerable reduction in measurement errors. Furthermore, the traffic monitoring device according to the invention advantageously enables the simultaneous detection of several vehicles.

[0008] The traffic monitoring device according to the invention can be designed as a part, in particular as a sub-assembly, of a traffic monitoring system, which can alternatively also be referred to as a speed measuring system or colloquially as a "speed camera". However, it would also be conceivable for the traffic monitoring device to constitute the entire traffic monitoring system.

[0009] The detection unit of the traffic monitoring device incorporates at least one event-based image sensor, but can also have multiple event-based image sensors, for example, two, three, four, or more. Event-based image sensors (EVS) differ fundamentally from conventional image sensors, such as conventional CMOS sensors with a fixed readout frequency (frame-based sensors (FBS)) or CCD sensors, which encode image brightness and generate a constant volume of data at a fixed frame rate, regardless of scene activity. In contrast, the pixels of the detection unit's event-based image sensor are each configured to independently and asynchronously detect relative changes in light intensity as events.Each pixel of the event-based image sensor is self-signaling and designed to respond individually, i.e., independently of other pixels of the event-based image sensor, and preferably in real time, to relative changes in light intensity. The pixels of the event-based image sensor are thus each configured as an independent photodetector. Examples of event-based image sensors, in contrast to the conventional image sensors mentioned above, are shown in "Activity-Driven, Event-Based Vision Sensors," T. Delbrück, et al., Proceedings of 2010 IEEE International Symposium on Circuits and Systems (ISCAS), pp. 2426-2429, in EP3424211A1, or in EP3732875A1. The design of the event-based image sensor is not limited to a specific type of photodetector.Pixels of the pixel matrix of the event-based image sensor can, but are not limited to, be designed as photodiodes and / or photocells and / or phototransistors and / or photoresistors and / or the like.

[0010] The pixel matrix of the event-based image sensor has a specific number of rows and columns, so that each pixel within the pixel matrix can be assigned unique coordinates. The pixel matrix is ​​not limited to a specific number of rows and / or columns. For example, currently available event-based image sensors have pixel matrices with 1280 columns and 720 rows. However, according to the invention, the pixel matrix of the event-based image sensor of the detection unit can also have a lower or higher number of rows and / or columns, and thus a lower or higher total number of pixels. This allows, in particular, flexible adaptation of the size and resolution of the event-based image sensor to different operating conditions, such as the width of the roadway to be monitored and / or the number of lanes to be monitored, and / or the like.

[0011] The event-based image sensor has a dynamic sampling rate, which is in particular at least 0.5 million events per second, advantageously at least 0.6 million events per second, particularly advantageously at least 0.7 million events per second, preferably at least 0.8 million events per second, preferably at least 0.9 million events per second and particularly preferably up to 1.0 million events per second.

[0012] Further information on the functionality of event-based image sensors can be found, for example, in WO 2019 / 129790 A1.

[0013] Preferably, sub-areas of the pixel matrix, and more preferably each individual pixel of the pixel matrix, are designed to be individually activatable and deactivatable. This advantageously allows for flexible adaptation to individual requirements. For example, individual sub-areas of an image area of ​​the event-based image sensor, which are not relevant for traffic monitoring, can be deactivated to increase the transmission rate. At the same time, this can also further reduce energy consumption and / or data volume. It is also conceivable to selectively deactivate sources of interference in sub-areas of the image area of ​​the event-based image sensor, for example, sources of regularly recurring changes in light intensity, such as warning lights, traffic lights, construction site lighting, and the like.

[0014] Since event detection using the event-based image sensor is not clocked, no time quantization takes place during the event detection itself. Therefore, the event-based image sensor preferably includes at least one timing unit, in particular at least one clock, for the precise detection of event occurrence times. Preferably, the clock has a temporal resolution of at least milliseconds, more preferably at least microseconds. The event-based image sensor preferably includes at least one quantization unit, which comprises the clock. The quantization unit is connected to the pixel matrix and is configured to convert events detected by the pixels and output by them as analog signals into digital signals and to assign each event at least the coordinates of the pixel by which it was detected, as well as a unique time.The quantization unit can also assign further parameters to the detected events, such as an absolute value of the detected relative change in light intensity and / or a light polarity, as a measure of whether the detected relative change in light intensity is an increase or decrease relative to an initial value, and / or similar parameters. The quantization unit can be part of a BUS controller for an interface of the event-based image sensor.

[0015] Theoretically, the event-based image sensor could be configured for direct alignment with the roadway without any additional optical instruments installed in front of it. Preferably, however, the detection unit has at least one, and in particular exactly one, lens which, when mounted, is positioned in front of the event-based image sensor in such a way that light can pass through the lens and strike the pixel matrix of the event-based image sensor. The lens is configured to determine the magnification of an image area of ​​the event-based image sensor as well as the size of a detectable image section. Depending on the requirements, lenses with different focal lengths and / or angles of view can therefore be used.If the detection unit has a lens, it is advisable to calibrate the lens's distortion correction when commissioning the traffic monitoring device, particularly to avoid errors in speed detection due to image distortions in the image area that may be caused by lens distortion. Furthermore, it is advisable to calibrate the distance between the event-based image sensor and detectable objects within its image area, at least during commissioning, in order to define and use as the basis for speed detection a clear relationship between the spatial distances of detected events in the image area and the spatial distances of real objects that trigger the detected events.

[0016] The distance between the event-based image sensor and detectable objects within its field of view could be determined once during the commissioning of the traffic monitoring device or at regular intervals during operation, for example, using a portable distance measuring device, stored, and used as the basis for speed measurement. However, in an advantageous embodiment of the invention, it is proposed that the detection unit include at least one distance sensor for measuring the distance between the event-based image sensor and the vehicle on the roadway. Such an embodiment advantageously improves the reliability of speed measurement. Furthermore, it advantageously increases ease of use, as manual distance measurement is no longer necessary.The distance sensor of the detection unit could, for example, be configured as an acoustic distance sensor, such as an ultrasonic sensor. Preferably, the distance sensor of the detection unit is configured as an optical distance sensor. In a preferred embodiment of the invention, it is proposed that the distance sensor be configured as a time-of-flight (TOF) laser distance sensor. This advantageously allows for particularly precise and reliable distance measurement. Furthermore, it advantageously enables simple calibration of the image area of ​​the event-based image sensor. The TOF laser distance sensor is configured to emit a pulsed laser beam, which can be detected in the image area of ​​the event-based image sensor in the form of events.The acronym "TOF" stands for "time of flight" and describes a time-of-flight method for distance measurement known to those skilled in the art, whereby the TOF laser distance sensor is designed to implement this method. As an alternative to a TOF laser distance sensor, the use of an FMCW laser distance sensor ("frequency modulation continuous wave") would also be conceivable. Another alternative to a TOF laser distance sensor is a different LIDAR-based distance sensor (LIDAR - Light Detection and Ranging), such as the aforementioned TOF laser distance sensor or a phase-modulated laser distance sensor. A further alternative to the TOF laser distance sensor is a radar-based or ultrasound-based distance sensor with a detection rate of at least 1000 detections per second (1000 samples / second).

[0017] In an advantageous embodiment of the invention, it is proposed that an optical axis of the event-based image sensor be aligned at least substantially congruently with a measuring axis of the distance sensor. This advantageously enables particularly precise calibration and thus highly reliable velocity measurement. The optical axis is an imaginary axis of symmetry of the event-based image sensor. If the event-based image sensor is used with a lens, the optical axis passes through a focal point of the lens. The measuring axis of the distance sensor is a straight line that passes through a starting point on the distance sensor and an endpoint on an object whose distance is to be measured.The requirement that the optical axis of the event-based image sensor is aligned "at least substantially congruently" with the measuring axis of the distance sensor means that the two axes are arranged parallel to each other, within standard tolerances, and at a distance of no more than 50 cm from each other, preferably 10 cm (10 centimeters) from each other, so that when viewed along a vertical line passing through both axes, only one of the two axes is visible. Larger distances between the two axes are conceivable, but this may reduce the measurement accuracy, particularly if the vehicle moves in such a way that the distance to the image sensor along the measuring axis changes during its passage.

[0018] Furthermore, it is proposed that the acquisition unit includes an evaluation unit connected to an interface of the event-based image sensor and configured to determine the vehicle's speed using at least one algorithm based on a temporal sequence of at least two detected events. This advantageously enables efficient determination of the vehicle's speed. The evaluation unit preferably comprises a digital signal processor on which the algorithm is stored in an executable form. Preferably, the evaluation unit includes a buffer memory, more preferably a first-in, first-out (FIFO) buffer memory, configured to temporarily store events detected by the event-based image sensor and transmitted via the interface in the temporal sequence of their occurrence and to transmit them to the digital signal processor for further processing.The interface of the event-based image sensor is preferably connected to and / or integrated into the quantization unit. Various bus systems known to those skilled in the art can be used as the interface for the event-based image sensor. The algorithm comprises at least one evaluation step in which the vehicle's speed is determined from the temporal sequence of at least two recorded events, for example, based on a distance-time calculation. The algorithm may include further steps, such as a detection step for identifying vehicles from a series of events and / or a verification step for checking the plausibility of the determined speed.Possibilities for determining the speed of at least one vehicle from event data captured by the event-based image sensor are the subject of the inventive method described below, wherein individual process steps of the method may be implemented in the algorithm.

[0019] It is further proposed that the detection unit be designed to simultaneously measure the speed of multiple vehicles in different lanes of the roadway. This can advantageously increase efficiency. Simultaneous speed measurement of multiple vehicles in different lanes proves particularly beneficial on roads with many lanes, such as highways. The detection unit could perform simultaneous speed measurement of at least two, four, or more vehicles. The different lanes could be arranged directly or indirectly adjacent to one another. Preferably, the detection unit is designed to simultaneously measure the speed of multiple vehicles in two directly adjacent lanes.However, it is also conceivable that the vehicles are located on lanes that are spaced apart from each other and are indirectly connected, for example by means of a bicycle lane, a section of vegetation or similar.

[0020] Furthermore, it is proposed that the detection unit be designed to measure the vehicle's speed regardless of its direction of travel on the road. This would advantageously increase the flexibility of the traffic monitoring device's deployment. For example, it would enable simultaneous monitoring of multiple lanes with partially opposing directions of travel. Additionally, it would be advantageous to also allow for the detection of wrong-way drivers.

[0021] Furthermore, it is proposed that the detection unit be designed as a mobile unit and include an energy storage device for off-grid power supply. This would advantageously enable particularly simple, flexible, and demand-oriented deployment of the traffic monitoring device at various traffic hotspots. The energy storage device can, but is not limited to, be designed as a rechargeable battery, for example, a lithium-ion battery or the like. Preferably, the traffic monitoring unit includes a support and / or stabilizing element, such as a tripod or the like, for temporary installation and adjustment of the detection unit. Alternatively, instead of a mobile unit, the detection unit could also be designed as a stationary installation.

[0022] In an advantageous embodiment of the invention, it is proposed that the detection unit be arranged in a mounted state next to the roadway, with an optical axis of the event-based image sensor being aligned at least substantially perpendicular to a longitudinal extent of a section of the roadway. Such an arrangement advantageously enables high measurement accuracy when the traffic monitoring device is positioned at various sections of roadways. This further increases the efficiency of the traffic monitoring device. In the mounted state, the detection unit is installed and adjusted, at least temporarily or permanently. "At least substantially perpendicular" here refers to an angle between 75° and 105°, preferably between 85° and 95°, and particularly advantageously 90°.The invention expressly also functions when the optical axis of the event-based image sensor is oriented in a way that is not at least substantially perpendicular to the longitudinal extent of a section of the roadway, but where a speed is determined that is lower than the actual speed of a vehicle. The section of the roadway could be essentially straight, such that its longitudinal extent corresponds to the longest edge of an imaginary geometric cuboid that just completely encloses the section. However, it is also conceivable that the section of the roadway is curved and its longitudinal extent runs parallel to a tangent at a vertex of the curve, with the optical axis then preferably oriented such that it intersects the vertex of the curve of the section.However, the use of the detection unit is not limited to a position next to the roadway. In its installed state, the unit could alternatively be positioned above the roadway, for example on a bridge, tunnel, or similar structure.

[0023] The invention further relates to a traffic monitoring system comprising at least one traffic monitoring device according to one of the previously described embodiments, at least one camera for capturing a vehicle's license plate, and a communication unit for wireless communication between the capture unit and the camera. Such a traffic monitoring system is characterized in particular by the advantageous properties of the traffic monitoring device described above. The traffic monitoring system could comprise several cameras, each of which could be directed at a lane of the roadway. The camera(s) can additionally include a flash. A flash illuminates at least the vehicle's license plate, but can also illuminate the vehicle and the surrounding scenery. A flash can also illuminate the vehicle interior and / or the driver.A flash unit can also be located at a distance from the camera, with the communication unit enabling wireless communication between the camera and the flash unit. Multiple flash units are also conceivable. Alternatively, the camera can have a built-in flash and optionally one or more flash units. The communication unit comprises at least one communication element connected to, preferably integrated into, the detection unit. The communication unit also comprises at least one further communication element connected to, preferably integrated into, the camera. The communication elements are configured for wireless, preferably bidirectional, communication with each other, which can be based on a suitable wireless communication standard, such as ISM 433 MHz, ISM 868 MHz, Bluetooth, BLE, WLAN, WPAN, infrared, or the like.In general, all ISM frequency bands (Industrial, Scientific, and Medical bands) suitable for wireless data transmission are appropriate. The detection unit is designed to send a trigger signal to the camera to capture the vehicle's license plate when it detects a vehicle exceeding a permitted speed limit. It is also conceivable that the images captured by the camera can be wirelessly transmitted via the communication unit to an external storage unit, for example, to provide redundant storage and / or to relieve the burden on the camera's local memory. The external storage unit can be connected to or integrated with the detection unit, or it can be part of an external data server connected to the communication unit via the internet.

[0024] The invention further relates to a method for determining the speed of at least one vehicle on a roadway based on the detection of relative changes in light intensity generated by the vehicle.

[0025] It is proposed that an event-based image sensor, comprising a pixel matrix with a plurality of pixels, be used to detect at least two relative changes in light intensity through at least two of the pixels as events, and that the vehicle's speed be determined from a temporal sequence of detected events. This advantageously provides a particularly efficient method for determining vehicle speed. Preferably, the method is carried out using the traffic monitoring device described above. The method particularly comprises at least one detection step in which events are recorded and at least one subsequent evaluation step in which the vehicle's speed is determined from the temporal sequence of recorded events.In the data acquisition step, each recorded event is assigned at least a time of recording and the coordinates of the pixel that recorded the respective event. In the evaluation step, vehicles can first be identified from the recorded events, for example, based on the number and temporal sequence of events in individual rows of the pixel matrix.In the evaluation step, the speed of events can then be determined based on their temporal sequence and their relative arrangement within an image area of ​​the event-based image sensor, in pixels per second. This speed is subsequently converted into the vehicle's speed in a desired unit, such as kilometers per hour or miles per hour. Conversion factors can be stored, for example, in a lookup table, particularly in the memory of the evaluation unit of the traffic monitoring device. Preferably, the evaluation step of the method is implemented in an algorithm stored in the evaluation unit of the traffic monitoring device.

[0026] Preferably, the method also includes a calibration step in which the image area of ​​the event-based image sensor is calibrated to the captured scene. In this calibration step, distances between pixels of the image area are calibrated to distances between objects within the scene that can be captured as events. Preferably, the calibration step precedes the capture step. In the calibration step, at least one distance measurement is performed between the event-based image sensor and at least one object within the scene. Preferably, the distance measurement is performed using a laser, in particular a pulsed laser, which is preferably generated by the distance sensor of the capture unit of the traffic monitoring device, and which distance measurement is used for calibrating the image area.The laser's measurement axis can be visible as a point or a line of individual events within the image area of ​​the event-based image sensor and can also be used for calibration, along with the distance measurement. If a lens is used that is positioned in front of the event-based image sensor, its focal length and angle of view are taken into account during the calibration of the image area. In addition, the lens's distortion correction is preferably also calibrated in this calibration step.

[0027] Based on these distance values, the corresponding calibration is then performed for converting the temporal sequence of events from pixels per second (pixel / s) to a speed in meters per second (m / s) or kilometers per hour (km / h).

[0028] Furthermore, it is proposed that the plausibility of the determined speed be verified by examining the temporal sequence and relative arrangement of the pixels whose recorded events form the basis for the speed determination. This can advantageously increase the reliability of the method. Preferably, the determined speed is verified in a verification step of the method, which may be sequential to the evaluation step. The plausibility of the determined speed can be verified, for example, by comparing the speeds of all events that move along a horizontal line of the image area during a given recording interval. This verification step can be implemented in the algorithm stored in the evaluation unit of the traffic monitoring device.

[0029] If the image sensor's edges are not horizontally aligned with the longitudinal extent of a section of the road, then all events that move across the event-based image sensor during a given recording interval will move along an oblique axis corresponding to the image sensor's orientation. In this case, this axis is calibrated to also capture the speed of events moving obliquely across the image sensor. Thus, in a computer program, the direction of event movement on the image sensor can be calibrated to determine the actual horizontal orientation within the scene. The image sensor's orientation is then used to calibrate the image sensor's pixels relative to the scene.

[0030] Individual steps of the procedure can be implemented in a computer program. The computer program can include instructions which, when executed by a computer, cause it to perform at least the evaluation step and / or the verification step of the procedure.

[0031] In an advantageous embodiment of the method, it is proposed that events occurring in adjacent pixels of the pixel matrix are grouped into related event groups. The movements of these event groups along an image area of ​​the event-based image sensor are tracked by detecting subsequent events that can be assigned to the same event group. The vehicle's speed is then determined from the speed of a movement of at least one event group along the image area. This embodiment advantageously enables a particularly reliable correlation between detected events and vehicles.An assignment of subsequent events to an event group can be achieved, for example, by comparing the magnitudes of the recorded relative changes in light intensity underlying the recorded events, as well as by means of a relative arrangement between the respective events and the respective subsequent events.

[0032] In an alternative, advantageous embodiment of the method, it is proposed that several regions of interest be defined within an image area of ​​the event-based image sensor, each of which is assigned a sub-area of ​​the pixel matrix. The vehicle's speed is then determined from the time interval between the occurrence of events in at least two regions of interest, which comprise different columns of the pixel matrix. This allows the method to be implemented particularly simply and with low computational effort.

[0033] Furthermore, it is proposed that the plausibility of the determined speed be verified by comparing a temporal sequence of events in at least two regions of interest, which comprise different rows of the pixel matrix. This can advantageously reduce, and preferably minimize, the susceptibility to errors. For example, if a difference in the time intervals between the occurrence of events in any two equidistant regions of interest in different rows of the pixel matrix is ​​detected, it can be concluded that these events are attributable to different objects, in particular vehicles, and the measurement can be discarded.If at least two successive events are detected in at least two regions of interest, which comprise different rows of the pixel matrix, the tilting of the front of a detected vehicle, for example due to a lane change, can also be detected. Accordingly, a determined speed can be corrected to take the detected tilt into account.

[0034] The traffic monitoring device, the traffic monitoring system, and the method are not to be limited to the applications and embodiments described above. In particular, the traffic monitoring device and / or the traffic monitoring system may, to achieve a functionality described herein, comprise a different number of individual elements, components, and / or units than specified herein. The method may comprise a different number of process steps than specified herein. Furthermore, for the value ranges specified in this document, values ​​within the stated limits are also to be considered disclosed and freely usable. Brief description of the drawings

[0035] The invention is explained in more detail below by way of example with reference to the drawings. The drawings show two embodiments of the invention. The drawings, the description, and the claims contain numerous features in combination. A person skilled in the art will expediently consider the features individually and combine them into meaningful further combinations. The drawings show: Fig. 1 a schematic representation of a traffic monitoring system with two cameras, a communication unit and a traffic monitoring device, which includes a detection unit for recording the speed of at least one vehicle on a roadway, Fig. 2 The detection unit with an event-based image sensor and a distance sensor in a schematic view, Fig. 3another schematic view of the acquisition unit with the event-based image sensor, which comprises a pixel matrix with a large number of pixels, and with an evaluation unit, which is connected to the event-based image sensor via an interface, Fig. 4 a schematic diagram illustrating a procedure for determining the speed of at least one vehicle on a roadway, Fig. 5 two schematic graphical representations of an image area of ​​the event-based image sensor for two successive time intervals to illustrate a first embodiment of the method, and Fig. 6 ,a schematic graphic representation of an image area of ​​the event-based image sensor in a time interval to illustrate a second embodiment of the method. Ways to implement the invention

[0036] Figure 1Figure 60 shows a traffic monitoring system 60 in a highly simplified and schematic top view. The traffic monitoring system 60 includes a traffic monitoring device 10.

[0037] The traffic monitoring device 10 comprises at least one detection unit 12 for detecting the speed of at least one vehicle 14 on a roadway 18 based on relative changes in light intensity generated by the vehicle 14. The detection unit 12 has at least one event-based image sensor 20. The event-based image sensor 20 comprises a pixel matrix 22 with a plurality of pixels 24, 26 (see figure). Figure 3 ), wherein pixels 24, 26 are each configured to register relative changes in light intensity independently and asynchronously as events 28, 30 (cf. Figures 5 and 6 ) to record.

[0038] The detection unit 12 is designed for the simultaneous speed measurement of several vehicles 14, 16 on different lanes 46, 48 of the roadway 18. Figure 1 shows, by way of example, vehicle 14 and another vehicle 16, with vehicle 14 moving on lane 46 and the other vehicle 16 on lane 48 of the roadway 18. However, the detection unit 12 is not limited to the simultaneous speed measurement of two vehicles 14, 16, but could also simultaneously measure the speeds of three, four, or more vehicles 14, 16, provided they are simultaneously within the field of view 86 of the event-based image sensor 20. Likewise, the detection unit could also simultaneously measure the speeds of two or more vehicles 14, 16 moving one behind the other on one of the lanes 44, 46, provided they are simultaneously within the field of view 86.Likewise, the number of two lanes 46, 48 of the carriageway 18 shown is merely exemplary and not to be understood as limiting for a simultaneous speed detection of vehicles 14, 16 by the detection unit 12.

[0039] In the Figure 1 Vehicle 14 is shown by way of example traveling in direction 42, and the other vehicle 16 is shown traveling in a further direction 44 opposite to direction 42. In this case, the detection unit 12 is designed to measure the speed of vehicle 14 on the roadway 18, regardless of its direction of travel 42. The detection unit 12 can simultaneously measure the speed of vehicle 14 traveling in direction 42 and the speed of the other vehicle 16 traveling in the opposite direction 44.

[0040] The detection unit 12 includes at least one distance sensor 36 (see Figure 2) for detecting the distance between the event-based image sensor 20 and the vehicle 14 and / or the other vehicle 16 on the roadway 18. In this case, the distance sensor 36 is designed as a TOF laser distance sensor. Alternatively, however, the use of other types of distance sensors known to those skilled in the art would also be conceivable without departing from the scope of the present invention.

[0041] An optical axis 38 of the event-based image sensor 20 is aligned at least substantially with a measuring axis 40 of the distance sensor 36 (see Figure 1).

[0042] The detection unit 12 is arranged in a mounted state next to the roadway 18, with the optical axis 38 of the event-based image sensor 20 being oriented at least substantially perpendicular to a longitudinal extent 52 of a subsection 54 of the roadway 18. In this case, the subsection 54 of the roadway is essentially straight. Alternatively, the detection unit 12 could also be arranged in the mounted state next to a curved subsection (not shown) of the roadway 18, with the optical axis 38 of the event-based image sensor 20 being oriented at least substantially perpendicular to a longitudinal extent of a curve of the subsection, wherein the longitudinal extent runs parallel to a tangent through a vertex of the curve of the subsection and the optical axis 38 intersects this vertex.

[0043] In the present embodiment, the detection unit 12 is designed as a mobile unit and includes an energy storage device 50 for independent power supply. The energy storage device 50 can, but is not limited to, be designed as a rechargeable battery. The detection unit 12 can, for example, be arranged on a tripod (not shown) of the traffic monitoring device 10 in order to align the optical axis 38 of the event-based image sensor 20 and the measuring axis 40 of the distance sensor 36 substantially parallel to a surface of the roadway 18. However, the detection unit 12 is not limited to use as a mobile unit and could alternatively also be designed as a stationary installation.Likewise, the detection unit 12 is not limited to an arrangement next to the carriageway 18, but could alternatively also be arranged temporarily or permanently above the carriageway, for example on a bridge or in a tunnel (not shown).

[0044] In addition to the traffic monitoring device 10, the traffic monitoring system 60 comprises at least one camera 56 for capturing the license plate of vehicle 14 and a communication unit 58 for wireless communication between the detection unit 12 and the camera 56. In this embodiment, the camera 56 is located next to the roadway 18 and is directed towards lane 46 of the roadway 18. The traffic monitoring system 60 also includes, in this embodiment, a further camera 84 for capturing the license plate of another vehicle 16, which is located next to the roadway 18 and is directed towards lane 48 of the roadway 18. The communication unit 58 comprises at least one communication element 88 which is connected to the receiving unit 12.The communication unit 58 also includes a further communication element 90, which is connected to the camera 56, and a further communication element 92, which is connected to the additional camera 84. In this case, the communication elements 88, 90, and 92 are configured for wireless bidirectional communication, i.e., for both wireless transmission and wireless reception of data, using a wireless communication standard such as Bluetooth, WLAN, or the like. If the speed of one of the vehicles 14 or 16, as detected by the sensor unit 12, exceeds a maximum speed permitted on section 54, the sensor unit 12 sends corresponding trigger signals to the camera 56 and / or the additional camera 84 for license plate recognition via the communication unit 58.

[0045] Figure 2Figure 1 shows the detection unit 12 of the traffic monitoring device 10 in a highly simplified schematic front view with the event-based image sensor 20 and the distance sensor 36. In this case, the distance sensor 36 is arranged above the event-based image sensor 20 in order to ensure at least substantially the same alignment between the measuring axis 40 of the distance sensor 36 and the optical axis 38 of the event-based image sensor 20 (see Figure 1). Figure 1 ) to enable. The detection unit has a lens 94, which is arranged in front of the event-based image sensor 20 such that light passes through the lens 94 onto the pixel matrix 22 (see figure). Figure 3 ) of the event-based image sensor 20. Before the traffic monitoring device 10 is put into operation, a distortion correction of the lens 94 is calibrated.

[0046] Figure 3Figure 1 shows the detection unit 12 of the traffic monitoring device 10 in a further highly simplified schematic view.

[0047] In the Figure 3 The pixel matrix 22 of the event-based image sensor 20 with the multitude of pixels 24, 26 is shown. For illustrative purposes and to simplify the representation, the pixel matrix 22 is shown in the Figure 3The matrix is ​​represented as an 8 x 8 matrix and has a total of 64 pixels, although for clarity, not every pixel is labeled with a reference symbol. For example, pixel matrix 22 has eight columns 76, 78 and eight rows 80, 82, again with the same omission of reference symbols. Based on columns 76, 78 and rows 80, 82 of pixel matrix 22, unique coordinates can be assigned to each of pixels 24, 26. For example, pixel 24 is located in column 76 and row 82, and pixel 26 is located in column 78 and row 80. However, pixel matrix 22 could also have a significantly higher or lower number of columns 76, 80 and / or rows 80, 82, and thus a different total number of pixels 24, 26, than shown in the example. Figure 3 shown, exhibit.

[0048] The acquisition unit 12 has an evaluation unit 32. The evaluation unit 32 is connected to an interface 34 of the event-based image sensor 20. The evaluation unit 32 is configured to determine the speed of the vehicle 14, 16 using at least one algorithm (see figure). Figure 1 ) based on a temporal sequence of at least two recorded events 28, 30 (cf. Figures 5 and 6 ) to determine.

[0049] The event-based image sensor in this case has a quantization unit 94, which is designed to process the events 28, 30 captured by the pixels 24, 26 (cf. Figures 5 and 6The quantization unit 94 converts the analog signals generated by the pixels into digital signals and transmits them to the evaluation unit 32 via interface 34. The quantization unit 94 includes a clock with a temporal resolution of at least milliseconds. If a relative change in light intensity detected by one of pixels 24 or 26 exceeds a predefined threshold, this is recorded as an event 28 or 30. The quantization unit 94 then determines the coordinates of the pixel 24 or 26 that detected the event 28 or 30, as well as the exact time of the event 28 or 30, and transmits this information to the evaluation unit 32.

[0050] The evaluation unit 32 comprises a digital signal processor and a buffer memory (not shown) which is connected upstream of the digital signal processor. The buffer memory is configured as a first-in, first-out (FIFO) buffer memory, so that events 28 and 30 can be processed by the digital signal processor in the order in which they occur.

[0051] Figure 4 shows a schematic process flow diagram to illustrate a procedure for determining the speed of at least one vehicle 14, 16 on a roadway 18 (cf. Figure 1 ) based on a detection of relative changes in light intensity generated by the vehicle 14, 16.

[0052] In the process, the event-based image sensor 20, which comprises the pixel matrix 22 with the plurality of pixels 22, 24 (see Figure 3), at least two relative changes in light intensity by at least two of the pixels 22, 24 are detected as events 28, 30 and the speed of the vehicle 14, 16 is determined from a temporal sequence of detected events 28, 30 (cf. Figures 5 and 6 ) determined.

[0053] The procedure can be carried out using the traffic monitoring device 10.

[0054] The procedure comprises at least one acquisition step 98, in which events 28, 30 are acquired by pixels 24, 26 of the pixel matrix 22. The procedure also comprises at least one evaluation step 100 following acquisition step 98, in which the speed of the vehicle 14, 16 is determined from the temporal sequence of the acquired events 28, 30, in particular by the evaluation unit 32 (see figure). Figure 3 ).

[0055] In the present case, the procedure also includes a verification step 102. In verification step 102, the plausibility of the determined speed is checked on the basis of the temporal sequence of the recorded events 28, 30 and on the basis of a mutual arrangement of those pixels 24, 26 to each other, whose recorded events 28, 30 are used as the basis for determining the speed.

[0056] Two possible ways to carry out the procedure are presented based on the Figures 5 and 6 explained in more detail.

[0057] Figure 5 Figure 1 shows two simplified graphical representations of events 28, 30, which occurred within an image area 66 of the event-based image sensor 20 at successive time intervals and which were detected by pixels 24, 26 of the pixel matrix 22. A left representation of the Figure 5 shows image area 66 during a first time interval. A right-hand representation of the Figure 5Figure 66 shows image area 66 during a second time interval, which follows the first time interval. These time intervals can each, for example, cover a period of 20 milliseconds. For clarity, not all events shown (28, 30) are labeled with reference symbols.

[0058] In a first possible embodiment of the method, events 28, 30, which occur in adjacent pixels 24, 36 of the pixel matrix 22, are grouped into related event groups 62, 64. For example, the events 28, 30, which were detected in acquisition step 98 of the method, are grouped together with a multitude of other events into an event group 62. Further events are grouped into another event group 64. The grouping into event groups takes place in evaluation step 100 of the method. Movements of the event groups 62, 64 along the image area 66 of the event-based image sensor 20 are tracked by detecting temporally subsequent events 28', 30' that can be assigned to an event group 62, 64. For example, two subsequent events 28', 30', which were detected in the second time interval by the event-based image sensor 20, can be assigned to event group 62.The assignment of the subsequent events 28', 30' to the event group 62 can be carried out, for example, by comparing the magnitudes of the recorded relative changes in light intensity underlying the events 28, 30, 28', 30' and by means of a relative arrangement between the respective events 28, 30 and the respective subsequent events 28', 30'.

[0059] The speed of vehicle 14, 16 is then determined from the speed of a movement of at least one of the event groups 62, 64 along the image area 66. As can be seen from the representations of the Figure 5As can be seen, event group 62 moved from a lower left section of image area 66 to a lower right section of image area 60 in the second time interval. Simultaneously, event group 64 moved from an upper right section of image area 66 in the first time interval to a middle upper section of image area 66 in the second time interval. In the illustrated embodiment, event group 62 could, for example, correspond to vehicle 14 and event group 64 to vehicle 16 (see figure). Figure 1) can be assigned. The speeds of the movements of the event groups 62, 64 along the image area 66 are first determined in the unit pixels per second in the evaluation step 100 of the procedure and then, taking into account the respective distance between the event-based image sensor 20 and the vehicle 14, 16, which can be recorded in the detection step 96, in particular by means of the distance sensor 36, converted into speeds of the vehicles 14, 16 in a desired unit, for example in kilometers per hour.

[0060] Figure 6 Figure 1 shows a simplified graphical representation of events 28, 30 occurring within the image area 66 of the event-based image sensor 20 in a time interval to illustrate a second possible embodiment of the method.

[0061] In the image area 66 of the event-based image sensor 20, several regions of interest 68, 70, 72, 74 are defined. Each region of interest 68, 70, 72, 74 is assigned a sub-area of ​​the pixel matrix 20, which can consist of one or more pixels 24, 26.

[0062] In evaluation step 100, the speed of the vehicle 14, 16 is determined from the time interval between the occurrence of events 28, 30 in at least two regions of interest 68, 70, which comprise different columns 76, 78 (see Figure 3) of the pixel matrix 22. In evaluation step 100 of the procedure, the speed of the vehicle 14, 16 can first be determined in pixels per second from the temporal sequence of the occurrence of events 28, 30 in the regions of interest 68, 70, and then, taking into account the respective distance between the event-based image sensor 20 and the vehicle 14, 16, converted into a speed in a desired unit, for example, kilometers per hour.

[0063] In verification step 102 of the procedure, the plausibility of the determined speed is assessed by comparing a temporal sequence of events 28, 30 in at least two regions of interest 68, 72, which have different lines 80, 82 (cf. Figure 3The pixel matrix 22 is examined. For example, if events 28 and 30 occur simultaneously in regions of interest 68 and 72, and subsequently in regions of interest 70 and 74, it can be concluded that these events 28 and 30 are attributable to the same vehicle 14 and 16. If the determined speed from the time interval between the occurrence of events 28 and 30 in regions of interest 68 and 70 matches a determined speed from the time interval between the occurrence of events 28 and 30 in regions of interest 72 and 74, the measurement can be considered plausible and valid.If, however, a difference arises in the time intervals between the occurrence of events 28, 30, which occur in regions of interest 68, 70 and regions of interest 72, 74, each comprising pixels 24, 26 from the same rows 80, 82 of the pixel matrix 22, this suggests that the recorded events were caused by different vehicles 14, 16 or by relative changes in light intensity not attributable to vehicles 14, 16. The measurement can then be discarded. If, in recording step 98, temporally successive events 28, 30 are recorded in at least two regions of interest 68, 72, which comprise different rows 80, 82 of the pixel matrix 22, this can also be interpreted in verification step 102 as indicating a tilting of the front of vehicle 14, 16, for example, due to a lane change.Accordingly, a speed recorded in evaluation step 100 can be corrected in verification step 102. Reference symbol list

[0064] 10 Traffic monitoring device 12 Detection unit 14 Vehicle 16 Vehicle 18 Roadway 20 Event-based image sensor 22 Pixel matrix 24 Pixel 26 Pixel 28 Event 30 Event 32 Evaluation unit 34 Interface 36 Distance sensor 38 Optical axis 40 Measuring axis 42 Direction of travel 44 Direction of travel 46 Lane 48 Lane 50 Energy storage 52 Longitudinal extent 54 Subsection 56 Camera 58 Communication unit 60 Traffic monitoring system 62 Event group 64 Event group 66 Image area 68 Region of interest 70 Region of interest 72 Region of interest 74 Region of interest 76 Column 78 Column 80 Row 82 Row 84 Additional camera 86 Field of view 88 Communication element 90 Communication element 92 Communication element 94 Objective 96 Quantization unit 98 Acquisition step 100 Evaluation step 102 Verification step

Claims

1. Traffic monitoring device (10) with at least one detection unit (12) for detecting the speed of at least one vehicle (14, 16) on a roadway (18) based on relative changes in light intensity generated by the vehicle (14, 16), characterized by the fact that the detection unit (12) has at least one event-based image sensor (20) which comprises a pixel matrix (22) with a plurality of pixels (24, 26), wherein the pixels (24, 26) are each configured to detect relative changes in light intensity independently and asynchronously as events (28, 30).

2. Traffic monitoring device (10) according to claim 1, characterized by the fact thatthe acquisition unit (12) has an evaluation unit (32) which is connected to an interface (34) of the event-based image sensor (20) and is designed to determine the speed of the vehicle (14, 16) by means of at least one algorithm based on a temporal sequence of at least two recorded events (28, 30).

3. Traffic monitoring device (10 according to claim 1 or 2, characterized by the fact that the detection unit (12) includes at least one distance sensor (36) for detecting a distance between the event-based image sensor (20) and the vehicle (14,16) on the roadway (18).

4. Traffic monitoring device (10) according to claim 3, characterized by the fact thatthe distance sensor (36) is designed as a LIDAR-based laser distance sensor, for example a TOF laser distance sensor or a phase-modulated laser distance sensor, or as a radar-based or ultrasound-based distance sensor with at least 1000 samples / second.

5. Traffic monitoring device (10 according to claim 3 or 4, characterized by the fact that an optical axis (38) of the event-based image sensor (20) is aligned at least substantially congruent with a measuring axis (40) of the distance sensor (36).

6. Traffic monitoring device (10) according to one of the preceding claims, characterized by the fact that the detection unit (12) is designed to detect the speed of the vehicle (14, 16) independently of its direction of travel (42, 44) on the roadway (18).

7. Traffic monitoring device (10) according to one of the preceding claims, characterized by the fact thatthe detection unit is designed for the simultaneous speed detection of several vehicles (14, 16) on different lanes (46, 48) of the roadway (18).

8. Traffic monitoring device (10) according to one of the preceding claims, characterized by the fact that the acquisition unit (12) is designed as a mobile unit and has an energy storage device (50) for off-grid power supply.

9. Traffic monitoring device (10) according to one of the preceding claims, characterized by the fact that the detection unit (12) is arranged in a mounted state next to the roadway (18), wherein an optical axis (38) of the event-based image sensor is aligned at least substantially perpendicular to a longitudinal extent (52) of a subsection (54) of the roadway (18).

10. Traffic monitoring system (60) comprising at least one traffic monitoring device (10) according to one of the preceding claims, comprising at least one camera (56) for capturing a license plate of the vehicle (14, 16) and comprising a communication unit (58) for wireless communication between the capture unit (12) and the camera (56) or the camera (56) and a flash unit.

11. Method for determining the speed of at least one vehicle (14, 16) on a roadway (18) based on the detection of relative changes in light intensity generated by the vehicle (14, 16), characterized by the fact thatby means of an event-based image sensor (20) comprising a pixel matrix (22) with a plurality of pixels (24, 26), at least two relative changes in light intensity by at least two of the pixels (24, 26) are detected as events (28, 30) and the speed of the vehicle (14, 16) is determined from a temporal sequence of detected events (28, 30).

12. Method according to claim 11, characterized by the fact that The plausibility of the determined speed is checked based on the temporal sequence and mutual arrangement of those pixels (24, 26) whose recorded events (28, 30) are used as the basis for determining the speed.

13. Method according to claim 11 or 12, characterized by the fact thatEvents (28, 30) occurring in adjacent pixels (24, 36) of the pixel matrix (22) are grouped into related event groups (62, 64), whose movements along an image area (66) of the event-based image sensor (20) are tracked by detecting temporally subsequent events (28', 30') that can be assigned to an event group (62, 64), wherein the speed of the vehicle (14, 16) is determined from a speed of a movement of at least one event group (62, 64) along the image area (66).

14. Method according to claim 11 or 12, characterized by the fact thatIn an image area (66) of the event-based image sensor, several regions of interest (68, 70, 72, 74) are defined, each of which is assigned a sub-area of ​​the pixel matrix (22), whereby the speed of the vehicle (14, 16) is determined from a time interval between an occurrence of events (28, 30) in at least two regions of interest (68, 70), which comprise different columns (76, 78) of the pixel matrix (22).

15. Method according to claim 14, characterized by the fact that The plausibility of the determined speed is checked by comparing a temporal sequence of events in at least two regions of interest (68, 72), which comprise different rows (80, 82) of the pixel matrix (22).

Images